There is a program called WORDS by one Whitaker, apparently no longer maintained since early 2000s (his home page has disappeared, too), but the copies are still floating around. I am using it all the time. It uses some semi-empirical rules to deduct all possible forms for a word. Personally I find it very useful, and even thinking about rewriting and maintaining it.
Distributed with the program is a file LISTALL which is a list of all word forms in the dictionary that can possibly be generated. A caveat here is that the dictionary is quite large, and contains medieval Latin words and forms. There are markings in the database, but in LISTALL the distinction is lost.
As far as the technique goes, you first grep the LISTALL¹ file, then copy interesting forms and paste then to WORDS as input. This way it is possible to understand the algorithm's "idea" how the possible form was produced. For example,
$ grep -P 'oo$' listall
aizoo
antoo
arctoo
Arctoo
arctoo
. . .
May of these are not verbs. It makes sense to copy the grep result and paste it to WORDS as input². Here's the output from pasting a few words into its console from the above list.
=>curotrophoo
curotrophoo ======== UNKNOWN
=>Eoo
eo.o N 2 1 DAT S M Always mostfreq
eo.o N 2 1 ABL S M Always mostfreq
Eous, Eoi N (2nd) M [XXXDX] lesser
morning star; Oriental, dweller in the east; one of the horses of the Sun;
eo.o ADJ 1 1 DAT S M POS Always mostfreq
eo.o ADJ 1 1 DAT S N POS Always mostfreq
eo.o ADJ 1 1 ABL S M POS Always mostfreq
eo.o ADJ 1 1 ABL S N POS Always mostfreq
Eous, Eoa, Eoum ADJ [XXXDX] lesser
eastern; of the dawn; belonging to/of/set in the morning;
=>epogdoo
epogdo.o ADJ 1 1 DAT S M POS Always mostfreq
epogdo.o ADJ 1 1 DAT S N POS Always mostfreq
epogdo.o ADJ 1 1 ABL S M POS Always mostfreq
epogdo.o ADJ 1 1 ABL S N POS Always mostfreq
epogdous, epogdoa, epogdoum ADJ [DDXES] Late uncommon
epogdo.o ADJ 2 6 DAT S C POS Always mostfreq
epogdo.o ADJ 2 6 ABL S C POS Always mostfreq
epogdoos, epogdoos, - ADJ [DDXES] Late uncommon
nine-eighths; (music); whole and eighth;
=>euroo
euro.o ADJ 1 1 DAT S M POS Always mostfreq
euro.o ADJ 1 1 DAT S N POS Always mostfreq
euro.o ADJ 1 1 ABL S M POS Always mostfreq
euro.o ADJ 1 1 ABL S N POS Always mostfreq
eurous, euroa, euroum ADJ [XXXDX] lesser
eastern;
So naturally none of these are verb forms, and one was not even understood by the parser. In fact, this simple example only found the same [re]boo and inchoo as the above link. But searches in the local file can be as complex as you fancy, not limited by a Web application interface.
ADDENDUM: Col. William A. Whitaker (USAF-Retired), who worked on the ADA language while at DARPA and wrote the WORDS program in his retirement, passed away in December 2010. As far as I can tell, nobody is supporting the original version at this time.
¹ The file is referred to as LISTALL in the documentation, but the actual file name as distributed is lowercase listall. I am keeping it consistent with the documentation, except in actual usage examples.
² The program accepts whitespace and newline as word separator, so that many words at once, a sentence, or even whole paragraph can be pasted to its input console.
